* Upgrade MKLDNN to v0.18-rc and fix issue caused by lib/lib64

Upgrade MKLDNN to v0.18-rc
Also fix the issue during upgrade

test=develop

* Rebase MKLDNN to rls-v0.18 branch

Some issues in v0.18-rc which caused INT8 conv op unit test failure was fixed
in rls-v0.18 branch

test=develop

* Upgrade MKLDNN from v0.18rc to formal v0.18 tag

test=develop

* Fix the windows compile issue.

test=develop

add analyzer_transformer_test

Avoid ce fails on windows.

test=develop

Make parent_idx a dispensable output for beam_search op to support models saved by older paddle version. (#16106)

test=develop

Add PaddleDType INT32 support

* Fix the node's sort issue when the graph is changed.

test=develop

* Clean code

test=develop

Four points for enhancing Paddle-TRT

add distribute fpn proposals op, test=develop

add if use static engine for trt
test=develop

test=develop

- Support setting attr in config
- Support setting dtype and initializer for input in config
test=develop

* polish the cast op doc

test=develop

* follow comments

test=develop

* fix api.spec

test=develop

* add parallel graph dist test=develop

* update test=develop

* update style test=develop

Add box decoder and assign

* Enable INT8 ReQuantize OP
test=develop

* Clean code
test=develop

* Add comments
test=develop

* Revert "Clean code"
test=develop

This reverts commit a7a49b8aa214f9730cb84e11ea96da564fe4b4d9.

* Modify requantize op test
test=develop

* fix requantize UT by moving public function to public test file.
test=develop

* Fix test fail due to file address change.
test=develop

* Change file address for requantize op.
test=develop

* refine recurrent_op
test=develop

* remove unnecessary code
test=develop

refine embeddingseqpool grad

* In some case, the input may have one than one negative value.
test=develop

* fix matmul bug
test=develop

* test=develop, reconstruct layer helper to fit imperative usage

* test=develop, fix import error on py35

* test=develop, fix rnn gradient error

* test=develop, delete test use code

* test=develop, remove helper from imperative usage

* test=develop, fix test_base_layer using new helper

* test=develop, reconstruct layerhelper for imperative mode

* test=develop, reconstruct layerhelper for imperative mode

* test=develop, fix bug

* test=develop, fix test failed bug

* test=develop, fix test failed bug

* test=develop, fix test failed bug

* test=develop, fix bug

* test=develop, polish code

refine SetCpuMathLibraryNumThreads

* remove the ignored from is_empty and less_than test=develop

* fix api spec test=develop

* fix the api spec test=develop

* test=develop

* unified PE and compiler
test=develop

* Polish code
test=develop

remove legacy function in ExecutionContext

Add spectral norm op

* add ir memory optimize test case for IfElse op, test=develop

* fix some unitttest failure by force using the python memory_optimize, test=develop

* tweak comments, test=develop

* fix unittest, test=develop

* fix unittest, test=develop

test=develop

test=develop

test=develop